espnet3.components.data.dataset_module.parse_dataset_reference_config

source

espnet3.components.data.dataset_module.parse_dataset_reference_config(config: Mapping[str, Any] | DictConfig)

Extract dataset source fields from one dataset config entry.

• Parameters:config – Dataset entry mapping. Expected keys are listed below.
  • data_src: optional dataset source reference.
  • data_src_args: optional constructor kwargs for Dataset.
• Returns: Tuple of normalized (data_src, data_src_args).
• Raises:
  • TypeError – If data_src is neither a string nor None.
  • ValueError – If data_src_args is incompatible with dict(...).

Notes

Only data_src_args is forwarded to the dataset constructor. Top-level config fields such as name and transform are ignored here.

Examples

>>> parse_dataset_reference_config(
...     {"data_src": "mini_an4/asr", "data_src_args": {"split": "train"}}
... )
('mini_an4/asr', {'split': 'train'})
>>> parse_dataset_reference_config({"data_src_args": {"split": "test"}})
(None, {'split': 'test'})